Atlantodentoplasty using the anterior retropharyngeal approach for treating irreducible atlantoaxial dislocation with atlantodental bony obstruction: a retrospective study

Ethics statement

All patients provided a written informed consent, and the current study was approved by the Medical Ethics Committee of Henan Provincial People’s Hospital (202400152).

Study design

We retrospectively analyzed patients diagnosed with irreducible atlantoaxial dislocation and atlantodental bony obstruction treated with anterior atlantodentoplasty combined with posterior fixation and fusion at our institution between March 2017 and June 2021. The inclusion criteria were as follows: (1) atlantoaxial dislocation; (2) inability to reduce preoperative hyperextension and hyperflexion on dynamic radiography; (3) skull traction with a weight of 1/5 of the body weight under general anesthesia; and (4) bony obstruction between the anterior arch of the atlas and the odontoid process in conditions such as atlantodental osteophyte and/or hyperplasia and hypertrophy of the anterior tubercle of the atlas or odontoid process.

In total, 26 patients with irreducible atlantoaxial dislocation with atlantodental bony obstruction were included in the analysis. Among them, six were men and 20 women, and they were aged 23–74 years (average: 49.8±11.8 years). All patients underwent atlas occipitalization and basilar invagination, and three patients had a history of foramen magnum decompression. There were 21 cases of simple atlantodental bony obstruction and five cases of atlantodental bony obstruction combined with hypertrophy of the anterior tubercle of the atlas. Three patients presented with a history of trauma, and 21 did not. Further, the trauma history of two patients was uncertain history. All patients underwent anterior retropharyngeal atlantodentoplasty, followed by posterior occipitocervical fixation and fusion, as shown in Tables 1 and 2.

Table 1

General data of the patients (n=26)

Table 2

Clinical data of the patients (n=26)

Treatment

Postoperative management

The patients were placed in the 30° reverse Trendelenburg position to eliminate pharyngeal edema. The tracheal cannula was removed only if the air leak test result was normal. The anterior and posterior drainage tubes were removed 48–72 hours after the surgery, and the anterior–posterior and lateral cervical radiograph, MRI, and cervical spine CT scan or CT angiography were reconducted within 1 week after the surgery.

The patients wore a neck collar for approximately 2 months postoperatively to achieve external fixation. The patients were followed-up at 3 months, 6 months, and 1 year after the surgery. The recovery of nerve function was evaluated based on clinical signs and symptoms. Bone graft fusion was evaluated via cervical radiography or CT scan, and recovery from spinal cord compression was assessed via MRI. Figs. 1–3 show a typical example.

Fig. 1

A 66-year-old female with irreducible atlantoaxial dislocation complicated with atlantodental bony obstruction. (A) Preoperative lateral X-ray indicated atlantoaxial dislocation. (B) The atlantoaxial dislocation was not reduced in preoperative lateral hyper-extension X-ray. (C) Preoperative sagittal reconstructive computed tomography of the cervical spine presented atlantoaxial dislocation complicated with atlantodental bony obstruction. (D) The upper cervical spinal cord was compressed by the superior posterior part of dens based on preoperative sagittal T2-weighted magnetic resonance imaging.

Fig. 2

(A) The atlantoaxial dislocation was not reduced under heavy weight (1/5 of body weight) skull traction. (B) Atlantodentoplasty was completed via the anterior retropharyngeal approach. The tip of the K-wire indicated the anterior arch of the atlas. (C) Posterior C0–C3 instrumentation was completed after anterior atlantodentoplasty.

Fig. 3

(A) Postoperative X-ray showed C0–C3 fixation by screw and rod. (B) Postoperative reconstructive sagittal computed tomography (CT) indicated an anatomical reduction of the upper cervical spine and sufficient autograft bone. The osteophyte in the atlantodental joint and partial anterior tubercle of the atlas were removed. (C) Postoperative magnetic resonance imaging indicated the release of upper cervical spinal cord compression. (D) Follow-up at 12 months after fixation confirmed solid fusion by CT scan.

Measurements

The surgical duration and blood loss volume of anterior retropharyngeal atlantodentoplasty and the whole surgery were recorded. Radiographic data such as the anterior atlantodental interval (ADI), space available for the cord (SAC) at the foramen magnum, clivus–canal angle (CCA), and cervical medullary angle (CMA), and clinical data including the Japanese Orthopedic Association (JOA) score before the surgery, at 1 week after the surgery, and during the latest follow-up (ADI not applicable) were examined. The O-C2 angle (OC2A) and the postoperative and preoperative OC2A (dOC2A) were measured and calculated. Further, the JOA recovery rate at the most recent follow-up was calculated. A >75% improvement in the JOA scores was considered excellent; 50%–74%, good; 25%–49%, fair; and <25%, poor. The fusion time of the grafted bone and the development of complications were examined. Atlantoaxial joint fusion was defined as continuous callus formation between the skull base and the posterior arch of the C2 on lateral radiograph. However, no motion was observed between the occiput and cervical spine on dynamic radiography or continuous callus formation between the occiput and the posterior arch of the C2 or bilateral lateral atlantoaxial joints on CT scan.

Statistical analysis

The IBM SPSS software ver. 20.0 (IBM Corp., Armonk, NY, USA) was used for the statistical analysis. Data with a normal distribution were expressed as mean±standard deviation. Preoperative and postoperative ADI and OC2A were compared using the paired t-test. Meanwhile, the SAC, CCA, CMA, and JOA were analyzed using repeated analysis of variation. The enumeration data were expressed as percentages, and a p-value of <0.05 indicated statistically significant differences.

Etiology and characteristics of atlantodental bony obstruction

Osteoarthritis and abnormal bone structure of the atlantodental joint are the common causes of atlantodental bony obstruction.

Osteoarthritis of the atlantoaxial joint is a non-infectious inflammatory disease characterized by the degeneration of the atlantoaxial articular cartilage, joint narrowing, osteophyte hyperplasia, and relaxation of the surrounding ligaments attributed to various lesions [12,13]. Its incidence rate is 5.4%–18.2% [5]. Osteoarthritis of the atlantoaxial joint is caused by various factors including trauma, infection, inflammatory disease, and atlantoaxial instability [4,14,15]. In our case series, all patients were diagnosed with basilar invagination (Goel type A) with atlantoaxial dislocation.

Atlantoaxial dislocation is a common cause of atlantoaxial osteoarthritis that mainly occurs in the lateral atlantoaxial joints, and it accounts for 14% of all atlantodental joint abnormalities [3]. The presence of osteophyte in the atlantoaxial joint may be caused by the secondary self-stabilizing mechanism after atlantoaxial instability or dislocation. The treatment strategies between the two types of atlantoaxial osteoarthritis differ [2,14]. In most cases, atlantoaxial dislocations in osteoarthritis of the lateral atlantoaxial joints can be satisfactorily reduced via posterior release. However, patients with atlantodental osteoarthritis often have bone abnormalities such as osteophyte hyperplasia of the atlantodental joint and/or hypertrophy of the anterior tubercle of the atlas and odontoid processes. The bony structures that block the reduction of the atlantoaxial joint are challenging to reduce via posterior release [7]. Previous reports have revealed that ventral compression of the spinal cord can be relieved via resection of the odontoid process or partial resection [9,16].

If there are abnormal bone structures, the hypertrophic anterior tubercle or odontoid process can lead to developmental spinal stenosis of the atlas [14]. Therefore, even if the ADI returns to normal, spinal cord compression still occurs. The bony obstruction must be removed to restore the spinal canal space [17].

Surgical methods for irreducible atlantoaxial dislocation

There are ongoing debates about the best treatment methods and surgical approaches for managing irreducible atlantoaxial dislocation [18]. Wang et al. [19] reported that this type of atlantoaxial dislocation was treated with transoral release, reduction, and posterior fixation and that satisfactory reduction was achieved. In 2004, Goel [20] first reported the treatment of irreducible atlantoaxial dislocation via posterior lateral mass joint release. Although the reduction effect was good, it did not meet the ideal standards described in the article. Smith et al. [21] have shown that the efficacy of reduction was questionable. With the progress in posterior techniques, the lateral mass-release technique has also improved. Shang et al. [22] used special spreaders to release the lateral mass joints, anterior contracted articular capsule, and ligament structure via distraction. Salunke et al. [23] utilized partial bone drilling on the lateral mass joint to eliminate the joint capsule and the ligament attachment at the front of the atlantoaxial joint. This approach has successful reduction outcomes for bony fusion dislocation via posterior release [24]. These improvements in the posterior techniques have led to fewer indications for anterior release. Deepak et al. [25] even called it the posterior era for the treatment of atlantoaxial dislocation.

In atlantoaxial dislocation with abnormal bony structure of the atlantodental joint, osteophyte hyperplasia and hypertrophy of the anterior tubercle of the atlas and the odontoid process in the atlantodental joint were observed. Due to obstruction of these bone structures, the bony structures that block atlantoaxial reduction via a simple posterior release are challenging to remove. Even if the posterior arch is decompressed, compression on the ventral side of the spinal cord cannot be relieved except when completing odontoid resection [7]. Some reports have assessed posterior odontoid process resection [26]. Nevertheless, owing to limitations in the visual field, it is difficult to achieve full resection under direct visualization when operating around the odontoid process using the posterior approach. Further, there is a risk of incomplete decompression. In addition, the stability of the upper cervical region is significantly affected after removing the odontoid process, which may increase the risk of internal fixation failure. When the atlantodental bony abnormalities are removed using the transoral approach, these bone structures cannot be viewed directly. Part of the anterior arch of the atlas or even the whole anterior arch should be removed before removing the atlantodental osteophytes and hypertrophic bones. Excessive bone removal affects local stability, and there is a risk of cerebrospinal fluid leakage. In transoral surgery, cerebrospinal fluid leakage is a catastrophic complication, and patients with this condition are at risk of intracranial infection.

Advantages and disadvantages of anterior retropharyngeal atlantodentoplasty

In 2009, Hao et al. [27] first reported the use of the anterior retropharyngeal approach for treating irreducible atlantoaxial dislocations. Although this approach is not exposed to direct visualization as a transoral release, it can also achieve a relatively sufficient soft tissue release, including the contracted tissue in front of the anterior arch of the atlas and axis, the anterior part of the bilateral lateral atlantoaxial capsule, and the alar ligament, thereby achieving a reduction effect similar to that of transoral release. Moreover, this approach involves an aseptic incision that is associated with a low incidence of infection [28]. In this series, none of the patients developed infection after anterior release. Therefore, oral preparations such as tooth cleaning before surgery were not required.

The anterior retropharyngeal approach has significant advantages in the anterior release of irreducible atlantoaxial dislocation. Osteophyte formation and bony hypertrophy of the atlantodental joint occur behind the anterior arch of the atlas. These bone structures cannot be viewed directly using the transoral approach. In cases involving basilar invagination, the anterior arch bone of the atlas or the whole anterior arch of the atlas must be removed before resecting the hypertrophic bone. Concurrently, owing to the angle of vision, the anterior retropharyngeal approach can directly view the osteophytes between the atlantodental joint and the hypertrophic anterior tubercle of the atlas and odontoid process. By removing these structures that block reduction, the normal structure of the atlantodental joint was retained, and local stability was not affected. While completing atlantodentoplasty, soft tissue release can be achieved simultaneously to remove the contracted longus cervical muscle, longus capitis muscle, articular capsule of the atlantoaxial joint, and ligaments around the atlantoaxial joint. Soft tissue release using this approach qualifies as the reduction standard for transoral release. In this series, the postoperative ADI, CCA, and CMA reached the standard of anatomical reduction, and the patients satisfactorily recovered their neurological function.

The current study had several limitations. It has a small sample size, and it used a retrospective design. When loosening the upper cervical region, the surgical field is relatively deep because the incision is far from the surgical site and is not as clear as in direct transoral visualization. Therefore, surgery with an endoscope or a microscope is preferred to obtain a better exposure [29].

Dysphagia is the most common complication of this procedure. Because the OC2A increased significantly after the surgery based on our data, the change in the postoperative OC2A might not be a main factor affecting short-term dysphagia after surgery [30]. The high incidence of dysphagia may be caused by a greater cervical exposure and continuous traction of the soft tissue during the procedure. Due to a higher cervical exposure, the superior laryngeal nerve must be pulled to the cranial side during the surgery, which may lead to a risk of injury to the superior laryngeal nerve. Three patients presented with superior laryngeal nerve injury in this group, of whom one had a relatively severe injury to the superior laryngeal nerve. The postoperative irritating cough worsened when drinking or eating and led to inhalation pneumonia. The cough was relieved after 1 month of feeding via a nasal tube. In addition, the posterior pharyngeal wall must be continuously pulled during surgery, and the incidence of dysphagia after surgery is higher than that after subaxial anterior cervical surgery. Retropharyngeal release leads to retropharyngeal wall edema, and it is associated with a risk of asphyxia. Hence, the patients’ respiratory status should be closely monitored postoperatively. The endotracheal tube was removed only if the air leakage test results were normal. If dyspnea occurs after endotracheal tube removal, emergency endotracheal intubation or tracheostomy should be conducted. As most of these patients undergo occipitocervical fusion, intubation may be challenging to perform. Therefore, an emergency tracheotomy should be carried out immediately to open the airway.